Error blocks and a 349-day nuisance wave change the GJ 163 signal test near 109 days

We test the known GJ 163 velocity signal near 109 days in 201 HARPS-RVBank version 02 nights. A conditional search fits known planets and includes two daily alias ranges. In the primary long-period range, residual sign tail fractions rise from 0.0002 with one-day blocks to 0.0600 with 100-day blocks and 0.0478 with annual blocks. A daily alias has a higher score. A prior 349-day nuisance wave weakens the long-period peak. Matched SERVAL and DRS long-range annual fractions are 0.2562 and 0.3776. Conditional recovery is only 22 to 29 percent at an injected amplitude of 1.66 m/s. The evidence depends on nuisance terms and noise assumptions. Nonlinear orbit shape is fixed during resampling, so these are not fully calibrated planet false-alarm probabilities. No planet is confirmed or excluded. Code, original aggregate evidence, input hashes and retrieval steps are included. External measurement rows remain outside the bundle.

astronomy · GJ 163 · HARPS · radial velocity · daily aliases · correlated noise · conditional bootstrap